paradance.pipeline.LogarithmPCAPipeline

class paradance.pipeline.LogarithmPCAPipeline(dataframe: DataFrame | None = None, config_path: str | None = None, n_trials: int = 200)

Pipeline for processing and optimizing PCA with logarithmic transformations.

This pipeline extends the BasePipeline class to implement a specific process for optimizing Principal Component Analysis (PCA) with logarithmic transformations, particularly focusing on self-balancing mechanisms.

file_type

Type of the file to load data from, supported types are ‘csv’ and ‘xlsx’.

Type:

str

dataframe

The loaded dataset in a pandas DataFrame.

Type:

pd.DataFrame

calculator

Calculator for PCA operations.

Type:

LogarithmPCACalculator

objective

The optimization objective.

Type:

MultipleObjective

__init__(dataframe: DataFrame | None = None, config_path: str | None = None, n_trials: int = 200) → None

Initializes the pipeline with configuration and trial settings.

Parameters:

• config_path (Optional[str]) – Path to the configuration file. Defaults to None.

• n_trials (int) – Number of optimization trials to perform. Defaults to 200.

Methods

__init__([dataframe, config_path, n_trials])	Initializes the pipeline with configuration and trial settings.
plot_logarithm_distributions()	Plots the logarithmic distributions of the dataset.
plot_self_balancing_projected_distribution(...)	This method updates the PCA weights in the calculator's PCA component and then plots the distribution based on these updated weights.
run()	Run the main execution flow of the pipeline.
show_results()	Displays the results of the optimization process.

plot_logarithm_distributions() → None

Plots the logarithmic distributions of the dataset.

plot_self_balancing_projected_distribution(pca_weights: ndarray) → None

This method updates the PCA weights in the calculator’s PCA component and then plots the distribution based on these updated weights.

show_results() → None

Displays the results of the optimization process.